Data processing method and device based on rail transit signal equipment

Abstract

The invention relates to the technical field of data processing, and provides a data processing method and device based on rail transit signal equipment, and the method comprises the steps: collecting the multi-dimensional analog data of the rail transit signal equipment in a current operation period; extracting multi-dimensional feature data from the multi-dimensional analog quantity data, and determining a health index corresponding to the rail transit signal equipment in the current operation period according to the multi-dimensional feature data; and inputting the health index corresponding to the current operation cycle and the health index sequences corresponding to the plurality of continuous calendar operation cycles into the target deep learning time sequence prediction model to obtain the remaining service life of the rail transit signal equipment. According to the method, comprehensive perception of the operation state of the equipment is realized through the multi-dimensional analog quantity data, the operation state of the equipment is converted into a quantifiable health characterization index through the health index, and the residual service life of the equipment is pre-judged by using the model, so that the health trend of the equipment is predicted in advance, and the fault risk is reduced.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a data processing method and apparatus based on rail transit signaling equipment. Background Technology

[0002] Current rail transit signaling equipment, such as turnouts, relies primarily on periodic preventative maintenance and reactive maintenance after a fault occurs. Some existing monitoring systems can collect single parameters such as current curves during turnout switching and trigger alarms by setting fixed thresholds, such as current exceeding a certain peak value or switching time exceeding a specified duration. However, when parameters trigger thresholds, faults often have already occurred or are about to occur, making true predictive maintenance impossible. Summary of the Invention

[0003] This invention provides a data processing method and apparatus based on rail transit signaling equipment to solve the technical problem in the prior art of data processing based on rail transit signaling equipment that there is a single parameter threshold alarm, and when the parameter triggers the threshold, the fault has often already occurred or is about to occur, making it impossible to achieve true predictive maintenance.

[0004] This invention provides a data processing method based on rail transit signaling equipment, comprising the following steps: Collect multi-dimensional analog data of rail transit signaling equipment during the current operating cycle; Multidimensional feature data is extracted from the multidimensional analog data, and the health index of the rail transit signaling equipment in the current operating cycle is determined based on the multidimensional feature data. The health index corresponding to the current operating cycle and the health index sequence corresponding to multiple consecutive historical operating cycles are input into the target deep learning time series prediction model to obtain the remaining service life of the rail transit signal equipment output by the target deep learning time series prediction model. The target deep learning time series prediction model is obtained by training an initial deep learning time series prediction model based on a health index time series containing multiple historical operating cycles and a remaining lifespan label corresponding to the health index time series.

[0005] According to the present invention, a data processing method based on rail transit signaling equipment is provided. The rail transit signaling equipment includes a turnout. The multidimensional analog data includes turnout electrical parameter monitoring data, turnout mechanical position status monitoring data, and vehicle speed data of trains passing the turnout. The method involves extracting multidimensional feature data from the multidimensional analog data and processing the data based on the multidimensional feature data, including: The starting peak value, switching time, and steady-state locking value are extracted from the turnout electrical parameter monitoring data. The turnout static range is determined based on the turnout electrical parameter monitoring data, and the static stable value is extracted from the turnout mechanical position status monitoring data based on the turnout static range. The starting peak value, the switching time, the steady-state locking value, the static stability value, and the vehicle speed data of the turnout are combined into multi-dimensional feature data.

[0006] According to a data processing method based on rail transit signaling equipment provided by the present invention, after determining the health index of the rail transit signaling equipment in the current operating cycle based on the multidimensional feature data, the method further includes: The time series trend chart of the health index, the turnout electrical parameter monitoring data chart, the turnout mechanical position status monitoring data chart, and the vehicle speed chart are overlaid on the same time axis; and... The feature space distribution map of the multidimensional feature data corresponding to each health index is superimposed and displayed in the same multidimensional coordinate system.

[0007] According to a data processing method for rail transit signaling equipment provided by the present invention, determining the health index of the rail transit signaling equipment in the current operating cycle based on the multidimensional feature data includes: The multidimensional feature data is mapped to data points in a preset multidimensional space; Determine the Mahalanobis distance between the data point and the health benchmark cluster in the preset multidimensional space; the health benchmark cluster is obtained by clustering sample data points of the rail transit signaling equipment in a healthy operating state; Based on the Mahalanobis distance, the health index of the rail transit signaling equipment corresponding to the current operating cycle is determined.

[0008] According to a data processing method based on rail transit signaling equipment provided by the present invention, after determining the health index of the rail transit signaling equipment corresponding to the current operating cycle, the method further includes: When the health index is lower than the alarm threshold, the multidimensional feature data is matched with a preset fault database to obtain a matching result; the preset fault database includes abnormal feature data corresponding to at least one preset fault mode of the rail transit signal equipment. Based on the matching results, the fault mode diagnosis results of the rail transit signaling equipment in the current operating cycle are determined.

[0009] According to a data processing method based on rail transit signaling equipment provided by the present invention, before extracting multidimensional feature data from the multidimensional analog data, the method further includes: The multidimensional analog data is filtered. The filtered multidimensional analog data is then subjected to Z-score normalization.

[0010] The present invention also provides a data processing device based on rail transit signaling equipment, comprising: The first data processing module is used to collect multi-dimensional analog data of rail transit signaling equipment during the current operating cycle; The second data processing module is used to extract multidimensional feature data from the multidimensional analog data, and determine the health index of the rail transit signaling equipment in the current operating cycle based on the multidimensional feature data. The third data processing module is used to input the health index corresponding to the current operating cycle and the health index sequence corresponding to multiple consecutive past operating cycles into the target deep learning time series prediction model to obtain the remaining service life of the rail transit signal equipment output by the target deep learning time series prediction model. The target deep learning time series prediction model is obtained by training an initial deep learning time series prediction model based on a health index time series containing multiple historical operating cycles and a remaining lifespan label corresponding to the health index time series.

[0011] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the program to implement the data processing method based on rail transit signaling equipment as described above.

[0012] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the data processing method based on rail transit signaling equipment as described above.

[0013] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the data processing method based on rail transit signaling equipment as described above.

[0014] The data processing method and apparatus for rail transit signaling equipment provided by this invention overcomes the limitations of traditional single-parameter monitoring by collecting multi-dimensional analog data within the current operating cycle of the rail transit signaling equipment, achieving comprehensive perception of the equipment's operating status. Based on the multi-dimensional analog data, multi-dimensional feature data is extracted and the health index corresponding to the current operating cycle is determined, transforming the complex operating status of the equipment into quantifiable and assessable health indicators that reflect the equipment's current health condition. Finally, the current health index and a sequence of health indices from multiple consecutive historical operating cycles are input into a target deep learning time series prediction model. Utilizing the model's deep analysis and prediction capabilities for time series data, the remaining service life of the equipment can be predicted, thereby enabling early prediction of equipment health trends and reducing the risk of failure. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating the data processing method based on rail transit signaling equipment provided in an embodiment of the present invention.

[0017] Figure 2 This is a schematic diagram of the structure of a data processing device based on rail transit signaling equipment provided in an embodiment of the present invention.

[0018] Figure 3 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0020] The data processing method based on rail transit signaling equipment in this embodiment of the invention, such as... Figure 1 As shown, it includes steps 110, 120 and 130.

[0021] Step 110: Collect multi-dimensional analog data of the rail transit signaling equipment during the current operating cycle.

[0022] It should be understood that rail transit signaling equipment refers to equipment used to ensure the safe operation of rail transit trains and control the train's path and speed, such as turnouts, signals, and track circuits. In this embodiment, turnouts are used as an example for explanation.

[0023] For turnout equipment, its operating cycle refers to the time interval for completing a complete switching action, that is, the entire process from the turnout receiving the switching command to the turnout switch rail and the stock rail being tightly fitted, the switch being in place, and the feedback signal being given. Under normal circumstances, the specific duration can be flexibly set based on the equipment model and track conditions.

[0024] Here, multidimensional analog data refers to continuously changing multidimensional physical quantity data that can reflect the operating status of the turnout. Typically, multidimensional analog data contains at least three or more key data types, such as the operating current of the turnout switching motor, the operating voltage during the switching process, the contact force between the switch rail and the stock rail, the switching time, and the motor housing temperature.

[0025] In one example, high-precision Hall current and voltage sensors can be installed in the power circuit of the turnout switching motor to collect motor current and voltage in real time; a laser displacement sensor can be installed at the actuating lever of the turnout switching mechanism to calculate the switching time based on displacement changes. All sensor signals are connected to an edge acquisition terminal, which communicates synchronously with the sensors via wireless communication technology. Acquisition is triggered when the turnout starts switching and stops after switching ends. The acquired data, including current, voltage, switching time, and motor temperature, are timestamped to form a multi-dimensional analog dataset for the current operating cycle, which is stored in a local database and uploaded to the central processing system.

[0026] Step 120: Extract multidimensional feature data from the multidimensional analog data, and determine the health index of the rail transit signaling equipment in the current operating cycle based on the multidimensional feature data.

[0027] Multidimensional feature data refers to the key features that can characterize the operating status of equipment extracted from multidimensional analog data through data processing and feature engineering. Each dimension of analog data corresponds to at least one feature index. For example, current data can extract features such as peak current, average current, current fluctuation variance, current rise time, and current fall time; voltage data can extract features such as average voltage, voltage fluctuation amplitude, and number of voltage drops, ultimately forming a multidimensional feature set.

[0028] The health index refers to a numerical indicator that reflects the current health status of equipment through quantitative calculation based on multi-dimensional feature data. The higher the health index, the more stable the current operating status of the equipment and the lower the potential risk of failure; the lower the health index, the more hidden dangers exist in the equipment and the higher the probability of failure.

[0029] In one example, the multidimensional analog data can be preprocessed first (e.g., outlier removal, noise smoothing, and data normalization). Then, feature engineering methods can be used to extract key features reflecting the equipment status from each dimension of the multidimensional analog data, forming multidimensional feature data. Next, based on the equipment's historical health data and failure cases, a correlation model between the multidimensional feature data and the health index is established. This correlation model then transforms the multidimensional feature data into a health index.

[0030] Step 130: Input the health index corresponding to the current operating cycle and the health index sequence corresponding to multiple consecutive past operating cycles into the target deep learning time series prediction model to obtain the remaining service life of the rail transit signaling equipment output by the target deep learning time series prediction model.

[0031] The target deep learning time series prediction model is obtained by training an initial deep learning time series prediction model based on a health index time series containing multiple historical operating cycles and a remaining lifespan label corresponding to the health index time series.

[0032] Here, remaining service life refers to the number of operating cycles from the end of the current operating cycle until the equipment fails to complete the conversion operation normally due to a malfunction.

[0033] It should be understood that the health index sequence corresponding to multiple consecutive historical operating cycles refers to an ordered data set composed of the health indices of the rail transit signaling equipment in N consecutive operating cycles prior to the current operating cycle, arranged in chronological order. For example, if the current cycle is the 100th operating cycle and N=30, then the historical health index sequence is a one-dimensional time series of the health indices formed in cycles 70-99.

[0034] In this embodiment, the health index sequence composed of the current health index and the historical continuous health index is used as input. The target deep learning time series prediction model is used to analyze the changing trend of the health index with the operating cycle. Combined with the correlation between the health index change and the remaining service life learned during the training of the target deep learning time series prediction model, the current remaining service life of the device is finally output.

[0035] In one example, a Long Short-Term Memory (LSTM) network can be selected as the initial deep learning time series prediction model. Using the past M years of operational data from this rail transit signaling equipment, multiple sets of historical health index time series and corresponding remaining service life labels are extracted: each time series represents the health index over N consecutive periods, and the remaining service life label represents the number of normal operating periods from the last period of that time series to the point before the rail transit signaling equipment malfunctions. The dataset is then divided into training and validation sets in a 7:3 ratio. Next, the initial deep learning time series prediction model is trained using mean squared error as the loss function and the Adam optimizer, iteratively trained for a preset number of epochs to obtain the target deep learning time series prediction model.

[0036] The data processing method based on rail transit signaling equipment provided in this invention overcomes the limitations of traditional single-parameter monitoring by collecting multi-dimensional analog data within the current operating cycle of the rail transit signaling equipment, achieving comprehensive perception of the equipment's operating status. Based on the multi-dimensional analog data, multi-dimensional feature data is extracted and a health index corresponding to the current operating cycle is determined, transforming the complex operating status of the equipment into quantifiable and assessable health indicators that reflect the equipment's current health condition. Finally, the current health index and a sequence of health indices from multiple consecutive historical operating cycles are input into a target deep learning time series prediction model. This allows the model to utilize its deep analysis and prediction capabilities for time series data to predict the remaining service life of the equipment, thereby enabling early prediction of equipment health trends and reducing the risk of failure.

[0037] In some embodiments, the rail transit signaling equipment includes a turnout, and the multidimensional analog data includes turnout electrical parameter monitoring data, turnout mechanical position status monitoring data, and vehicle speed data of trains passing the turnout. The step of extracting multidimensional feature data from the multidimensional analog data and, based on the multidimensional feature data, includes: The starting peak value, switching time, and steady-state locking value are extracted from the turnout electrical parameter monitoring data. The turnout static range is determined based on the turnout electrical parameter monitoring data, and the static stable value is extracted from the turnout mechanical position status monitoring data based on the turnout static range. The starting peak value, the switching time, the steady-state locking value, the static stability value, and the vehicle speed data of the turnout are combined into multi-dimensional feature data.

[0038] It should be noted that turnout electrical parameter monitoring data refers to continuous analog data reflecting the electrical operating status of the turnout's switching motor, and is a key basis for judging whether the turnout's electrical performance is normal. Examples include the turnout switching motor's operating current, operating power, and supply voltage data. The acquisition of turnout electrical parameter monitoring data must cover the turnout's start-up, switching, and locking phases.

[0039] Turnout mechanical position status monitoring data refers to continuous analog data collected by position sensing devices that reflects the spatial position and coordination status of the turnout mechanical structure. Examples include turnout gap value and turnout offset. The collection of turnout mechanical position status monitoring data needs to be synchronized with the data timestamp of turnout electrical parameter monitoring data collection.

[0040] The vehicle speed data when a train passes through a turnout refers to the real-time continuous analog data of the train's operating speed when passing through the turnout area. It is collected by speed sensors installed on the tracks before and after the turnout and reflects the dynamic impact load of the train on the turnout.

[0041] In one example, the starting peak value, transition time, and steady-state locking value are extracted from the working current or power data of the turnout electrical parameter monitoring data. For example, the time period during which the current rises from 0 to the first local maximum value is used, and the maximum current value within this stage is taken as the starting peak value. Then, taking the current starting moment as the starting point and the moment when the current drops to the steady-state value as the ending point, the time difference between the starting point and the ending point is taken as the transition time. Finally, the locking stage in the current data is identified (i.e., the time period during which the current fluctuates stably after the transition, and the fluctuation amplitude does not exceed the preset amplitude), and the average current value in this stage is calculated as the steady-state locking value.

[0042] Furthermore, the operating current data in the turnout electrical parameter monitoring data can be analyzed, and the time period in which the current is 0 or maintains a steady-state locking value for P consecutive seconds with fluctuations not exceeding a preset fluctuation range can be defined as the static interval. Then, all gap data collected within this static interval can be extracted from the turnout mechanical position status monitoring data, and the average value of all gap data collected within this static interval can be calculated as the static stable value. For example, the turnout gap value data and turnout offset data collected within this static interval can be extracted, and the average value of the turnout gap value data and turnout offset data collected within this static interval can be calculated as the static stable value.

[0043] Finally, the extracted start-up peak, transition time, steady-state locking value, static stability value, and vehicle speed data of trains passing through turnouts are combined in a fixed-dimensional order to form a multi-dimensional feature data set containing 5 feature dimensions, which is used for subsequent health index calculation.

[0044] The data processing method based on rail transit signaling equipment provided in this invention first extracts the starting peak value, switching time, and steady-state locking value, which directly reflect the status of the motor and power supply system, from the turnout electrical parameter monitoring data to capture potential fault signals at the electrical level. Second, it locates the static interval through the turnout electrical parameter monitoring data, eliminates the interference of switching actions on the mechanical state, and extracts the static stability value, which reflects static stability, from the turnout mechanical position status monitoring data. Finally, it combines the vehicle speed data of trains passing through the turnout and integrates the impact of external loads on turnout losses, ultimately forming multi-dimensional feature data covering electrical, mechanical, and external environmental aspects. This provides comprehensive multi-dimensional feature data for subsequent health index calculations, effectively overcoming the limitation that single-dimensional features cannot reflect the overall status of the turnout.

[0045] In some embodiments, after determining the health index of the rail transit signaling equipment corresponding to the current operating cycle based on the multidimensional feature data, the method further includes: The time series trend chart of the health index, the turnout electrical parameter monitoring data chart, the turnout mechanical position status monitoring data chart, and the vehicle speed chart are overlaid on the same time axis; and... The feature space distribution map of the multidimensional feature data corresponding to each health index is superimposed and displayed in the same multidimensional coordinate system.

[0046] It should be understood that the time series trend chart of the health index refers to a line chart or curve plotted with the operating cycle as the horizontal axis and the health index as the vertical axis, which plots the current and previous operating cycles in chronological order. It can intuitively show the trend of the health index changing with the cycle. The vertical axis can also usually mark the health threshold range, which makes it easy to quickly identify the nodes of deterioration of health status.

[0047] The turnout electrical parameter monitoring data graph corresponding to the health index refers to a line graph or curve plotted with the operating cycle corresponding to the health index as the horizontal axis and the turnout electrical parameter monitoring data as the vertical axis, and the turnout electrical parameter monitoring data of the current and continuous past operating cycles in chronological order. It can intuitively show the trend of the turnout electrical parameter monitoring data corresponding to each health index changing with the cycle, and realize the direct correlation and traceability between the health index and the original electrical parameters.

[0048] The turnout mechanical position status monitoring data graph corresponding to the health index refers to a line graph or curve plotted in chronological order, with the operating cycle corresponding to the health index as the horizontal axis and the turnout mechanical position status monitoring data as the vertical axis. It can intuitively present the trend of the turnout mechanical position status monitoring data corresponding to each health index changing with the cycle, and reflect the impact of mechanical status on the health index.

[0049] The vehicle speed graph corresponding to the health index refers to a line graph or curve plotted in chronological order, with the operating cycle corresponding to the health index as the horizontal axis and the speed of the train when passing through the turnout within this operating cycle as the vertical axis. It can reflect the relationship between external load and health index.

[0050] In one example, a multi-sub-graph linked timeline page can be constructed. The page contains four sub-graphs: the top shows the time series trend of the health index; below are the turnout electrical parameter monitoring data, the turnout mechanical position status monitoring data, and the vehicle speed data, respectively. All sub-graphs maintain synchronized time on their horizontal axes. When maintenance personnel click on the health index point for the corresponding operating period in the time series trend of the health index, the four sub-graphs below automatically switch to the corresponding data for that operating period.

[0051] Furthermore, in addition to overlaying all data on the same time axis, this embodiment can also display the multidimensional feature data of each running cycle as data points in the feature space distribution in a multidimensional coordinate system.

[0052] Specifically, a multidimensional coordinate system is a spatial coordinate system constructed based on the dimensions of multidimensional feature data. The number of dimensions matches the number of feature dimensions, and it is used to present the distribution patterns of feature data in space. Then, data points corresponding to multidimensional feature data of different health states can be presented in the multidimensional coordinate system using different colors. For example, based on the numerical range of the health index, the health index can be divided into multiple health states: healthy, sub-healthy, and risky. Data points corresponding to the healthy state are then presented in green, those corresponding to the sub-healthy state in yellow, and those corresponding to the risky state in red. Based on this rendering, the correlation and clustering patterns between feature data and health states can be intuitively presented.

[0053] The data processing method based on rail transit signaling equipment provided in this invention, on the one hand, overlays the time series change trend graph of the health index with the corresponding turnout electrical parameter monitoring data graph, turnout mechanical position status monitoring data graph, and vehicle speed graph on the same time axis, which facilitates maintenance personnel to trace the reasons for changes in the health index; on the other hand, it displays the spatial distribution of feature data in a multi-dimensional coordinate system, which can intuitively present the correlation and clustering rules between feature data and health status, effectively improving the efficiency of fault location and risk prediction.

[0054] In some embodiments, determining the health index of the rail transit signaling equipment in the current operating cycle based on the multidimensional feature data includes: The multidimensional feature data is mapped to data points in a preset multidimensional space; Determine the Mahalanobis distance between the data point and the health benchmark cluster in the preset multidimensional space; the health benchmark cluster is obtained by clustering sample data points of the rail transit signaling equipment in a healthy operating state; Based on the Mahalanobis distance, the health index of the rail transit signaling equipment corresponding to the current operating cycle is determined.

[0055] Here, the pre-defined multidimensional space refers to a mathematical space constructed with each feature dimension of the multidimensional feature data as the coordinate axis. Each coordinate axis corresponds to a feature parameter, and each point in the space represents a combination of multidimensional feature data for a certain operating cycle.

[0056] A health baseline cluster refers to a high-density region or set of central points formed by clustering a large number of equipment sample data points in a predefined multi-dimensional space, all of which are in a healthy operating state. The health baseline cluster represents the characteristic distribution range of rail transit signaling equipment in a healthy state.

[0057] In this embodiment, historical multidimensional feature data of the rail transit signaling equipment under known health conditions can be collected in advance for multiple historical operating cycles. Each historical operating cycle's historical multidimensional feature data is mapped to a sample data point. The K-means clustering algorithm is then used to cluster these sample data points, forming a health baseline cluster. The center point of this health baseline cluster and the covariance matrix of all sample points within the health baseline cluster to this center point are then determined.

[0058] Based on this, after obtaining the multidimensional feature data for each operating cycle, it is mapped to data points in a preset multidimensional space. Then, according to the center point of the health benchmark cluster and the covariance matrix of all sample points within the health benchmark cluster to this center point, the Mahalanobis distance from the current data point to this health benchmark cluster is calculated. Here, the calculation function for the Mahalanobis distance is consistent with the calculation function in the prior art, and will not be described in detail here.

[0059] In this embodiment, a Mahalanobis distance threshold can be preset. For example, historical multidimensional feature data from multiple operating cycles prior to the failure of the rail transit signaling equipment can be collected, and the Mahalanobis distance to the health baseline cluster can be calculated for each data point. The average value of these distances can then be used as the Mahalanobis distance threshold. Based on this, the health index is set to M (a preset maximum index value) when the Mahalanobis distance is zero, and to 0 when the Mahalanobis distance is not less than the Mahalanobis distance threshold. For Mahalanobis distances between 0 and the threshold, a preset conversion function is used to convert them into a health index between 0 and M. For example, HI = M * (1 - MD / T), where HI represents the health index, MD represents the Mahalanobis distance, M represents the preset maximum health index value, and T represents the preset Mahalanobis distance threshold.

[0060] The data processing method based on rail transit signaling equipment provided in this invention maps the multidimensional feature data of the current operating cycle to a data point in space; then calculates the Mahalanobis distance between the data point and the health benchmark cluster; finally, quantifies the current health status of the equipment based on the magnitude of the Mahalanobis distance. This method comprehensively considers the synergistic effects of all feature dimensions, avoids the limitations of simple threshold judgment, and achieves a refined and quantitative assessment of the health status of the equipment.

[0061] In some embodiments, after determining the health index of the rail transit signaling equipment corresponding to the current operating cycle, the method further includes: When the health index is lower than the alarm threshold, the multidimensional feature data is matched with a preset fault database to obtain a matching result; the preset fault database includes abnormal feature data corresponding to at least one preset fault mode of the rail transit signal equipment. Based on the matching results, the fault mode diagnosis results of the rail transit signaling equipment in the current operating cycle are determined.

[0062] Here, the alarm threshold refers to a critical value set within the range of the health index, used to trigger the fault diagnosis process. When the health index falls below this value, it indicates that the equipment is in an abnormal state, requiring further analysis of the specific cause of the fault.

[0063] In this embodiment, a preset fault database is set up in advance. This preset fault database includes various fault types that occur in the historical operation cycle of rail transit signal equipment, as well as abnormal feature data associated with each fault type.

[0064] Furthermore, when the health index of the current operating cycle is detected to be lower than the alarm threshold, the abnormal feature data corresponding to each preset fault mode in the preset fault database is traversed, and the matching degree between the abnormal feature data and the current multidimensional feature data is calculated. Then, the preset fault mode corresponding to the highest matching degree is used as the fault mode diagnosis result corresponding to the current operating cycle.

[0065] The data processing method based on rail transit signaling equipment provided in this embodiment of the invention comprehensively compares the multidimensional feature data of the current operating cycle with various preset fault modes and their abnormal feature data stored in a preset fault database; finally, based on the preset fault mode with the highest similarity, the fault mode diagnosis result corresponding to the current operating cycle is output, thereby improving the efficiency of fault diagnosis.

[0066] In some embodiments, before extracting multidimensional feature data from the multidimensional analog data, the method further includes: The multidimensional analog data is filtered. The filtered multidimensional analog data is then subjected to Z-score normalization.

[0067] In this embodiment, the acquired raw multidimensional analog data is first filtered to remove irrelevant high-frequency interference signals and effectively extract the valid signals. Next, based on the filtering process, the data is Z-score standardized to convert analog data with different parameter dimensions to a unified scale, eliminating the influence of differences in units and numerical ranges of different parameters.

[0068] The data processing apparatus based on rail transit signaling equipment provided by the present invention is described below. The data processing apparatus based on rail transit signaling equipment described below and the data processing method based on rail transit signaling equipment described above can be referred to in correspondence.

[0069] The data processing device based on rail transit signaling equipment in this embodiment of the invention, such as... Figure 2 As shown, it includes the following modules: The first data processing module 210 is used to collect multi-dimensional analog data of rail transit signaling equipment during the current operating cycle; The second data processing module 220 is used to extract multidimensional feature data from the multidimensional analog data, and determine the health index of the rail transit signaling equipment in the current operating cycle based on the multidimensional feature data. The third data processing module 230 is used to input the health index corresponding to the current operating cycle and the health index sequence corresponding to multiple consecutive past operating cycles into the target deep learning time series prediction model to obtain the remaining service life of the rail transit signal equipment output by the target deep learning time series prediction model. The target deep learning time series prediction model is obtained by training an initial deep learning time series prediction model based on a health index time series containing multiple historical operating cycles and a remaining lifespan label corresponding to the health index time series.

[0070] This embodiment of the data processing device based on rail transit signaling equipment overcomes the limitations of traditional single-parameter monitoring by collecting multi-dimensional analog data within the current operating cycle of the rail transit signaling equipment, achieving comprehensive perception of the equipment's operating status. Based on the multi-dimensional analog data, it extracts multi-dimensional feature data and determines the health index corresponding to the current operating cycle, transforming the complex operating status of the equipment into quantifiable and assessable health indicators that reflect the equipment's current health condition. Finally, the current health index and a sequence of health indices from multiple consecutive historical operating cycles are input into a target deep learning time series prediction model. Utilizing the model's deep analysis and prediction capabilities for time series data, it enables the prediction of the equipment's remaining service life, thereby achieving early prediction of equipment health trends and reducing the risk of failure.

[0071] Figure 3 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 3 As shown, the electronic device may include: a processor 310, a communications interface 320, a memory 330, and a communication bus 340, wherein the processor 310, the communications interface 320, and the memory 330 communicate with each other via the communication bus 340. The processor 310 can call logical instructions in the memory 330 to execute a data processing method based on rail transit signaling equipment, the method including: Collect multi-dimensional analog data of rail transit signaling equipment during the current operating cycle; Multidimensional feature data is extracted from the multidimensional analog data, and the health index of the rail transit signaling equipment in the current operating cycle is determined based on the multidimensional feature data. The health index corresponding to the current operating cycle and the health index sequence corresponding to multiple consecutive historical operating cycles are input into the target deep learning time series prediction model to obtain the remaining service life of the rail transit signal equipment output by the target deep learning time series prediction model. The target deep learning time series prediction model is obtained by training an initial deep learning time series prediction model based on a health index time series containing multiple historical operating cycles and a remaining lifespan label corresponding to the health index time series.

[0072] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc., each of which can store program code.

[0073] On the other hand, the present invention also provides a computer program product, the computer program product comprising a computer program that can be stored on a non-transitory computer-readable storage medium, wherein when the computer program is executed by a processor, the computer is capable of executing the data processing method based on rail transit signaling equipment provided by each of the above methods, the method comprising: Collect multi-dimensional analog data of rail transit signaling equipment during the current operating cycle; Multidimensional feature data is extracted from the multidimensional analog data, and the health index of the rail transit signaling equipment in the current operating cycle is determined based on the multidimensional feature data. The health index corresponding to the current operating cycle and the health index sequence corresponding to multiple consecutive historical operating cycles are input into the target deep learning time series prediction model to obtain the remaining service life of the rail transit signal equipment output by the target deep learning time series prediction model. The target deep learning time series prediction model is obtained by training an initial deep learning time series prediction model based on a health index time series containing multiple historical operating cycles and a remaining lifespan label corresponding to the health index time series.

[0074] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the data processing method based on rail transit signaling equipment provided by each of the above methods, the method comprising: Collect multi-dimensional analog data of rail transit signaling equipment during the current operating cycle; Multidimensional feature data is extracted from the multidimensional analog data, and the health index of the rail transit signaling equipment in the current operating cycle is determined based on the multidimensional feature data. The health index corresponding to the current operating cycle and the health index sequence corresponding to multiple consecutive historical operating cycles are input into the target deep learning time series prediction model to obtain the remaining service life of the rail transit signal equipment output by the target deep learning time series prediction model. The target deep learning time series prediction model is obtained by training an initial deep learning time series prediction model based on a health index time series containing multiple historical operating cycles and a remaining lifespan label corresponding to the health index time series.

[0075] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0076] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in each of the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of each embodiment of the present invention.

Claims

1. A data processing method based on rail transit signaling equipment, characterized in that, include: Collect multi-dimensional analog data of rail transit signaling equipment during the current operating cycle; Multidimensional feature data is extracted from the multidimensional analog data, and the health index of the rail transit signaling equipment in the current operating cycle is determined based on the multidimensional feature data. The health index corresponding to the current operating cycle and the health index sequence corresponding to multiple consecutive historical operating cycles are input into the target deep learning time series prediction model to obtain the remaining service life of the rail transit signal equipment output by the target deep learning time series prediction model. The target deep learning time series prediction model is obtained by training an initial deep learning time series prediction model based on a health index time series containing multiple historical operating cycles and a remaining lifespan label corresponding to the health index time series.

2. The data processing method based on rail transit signaling equipment according to claim 1, characterized in that, The rail transit signaling equipment includes turnouts, and the multidimensional analog data includes turnout electrical parameter monitoring data, turnout mechanical position status monitoring data, and vehicle speed data of trains passing the turnouts. The extraction of multidimensional feature data from the multidimensional analog data, and the determination of the multidimensional feature data based on the multidimensional feature data, includes: The starting peak value, switching time, and steady-state locking value are extracted from the turnout electrical parameter monitoring data. The turnout static range is determined based on the turnout electrical parameter monitoring data, and the static stable value is extracted from the turnout mechanical position status monitoring data based on the turnout static range. The starting peak value, the switching time, the steady-state locking value, the static stability value, and the vehicle speed data of the turnout are combined into multi-dimensional feature data.

3. The data processing method based on rail transit signaling equipment according to claim 2, characterized in that, After determining the health index of the rail transit signaling equipment in the current operating cycle based on the multidimensional feature data, the method further includes: The time series trend chart of the health index, the turnout electrical parameter monitoring data chart, the turnout mechanical position status monitoring data chart, and the vehicle speed chart are overlaid on the same time axis; and... The feature space distribution map of the multidimensional feature data corresponding to each health index is superimposed and displayed in the same multidimensional coordinate system.

4. The data processing method based on rail transit signaling equipment according to claim 1, characterized in that, The step of determining the health index of the rail transit signaling equipment in the current operating cycle based on the multidimensional feature data includes: The multidimensional feature data is mapped to data points in a preset multidimensional space; Determine the Mahalanobis distance between the data point and the health benchmark cluster in the preset multidimensional space; the health benchmark cluster is obtained by clustering sample data points of the rail transit signaling equipment in a healthy operating state; Based on the Mahalanobis distance, the health index of the rail transit signaling equipment corresponding to the current operating cycle is determined.

5. The data processing method based on rail transit signaling equipment according to claim 1, characterized in that, After determining the health index of the rail transit signaling equipment in the current operating cycle, the method further includes: When the health index is lower than the alarm threshold, the multidimensional feature data is matched with a preset fault database to obtain a matching result; the preset fault database includes abnormal feature data corresponding to at least one preset fault mode of the rail transit signal equipment. Based on the matching results, the fault mode diagnosis results of the rail transit signaling equipment in the current operating cycle are determined.

6. The data processing method based on rail transit signaling equipment according to claim 1, characterized in that, Before extracting multidimensional feature data from the multidimensional analog data, the process also includes: The multidimensional analog data is filtered. The filtered multidimensional analog data is then subjected to Z-score normalization.

7. A data processing device based on rail transit signaling equipment, characterized in that, include: The first data processing module is used to collect multi-dimensional analog data of rail transit signaling equipment during the current operating cycle; The second data processing module is used to extract multidimensional feature data from the multidimensional analog data, and determine the health index of the rail transit signaling equipment in the current operating cycle based on the multidimensional feature data. The third data processing module is used to input the health index corresponding to the current operating cycle and the health index sequence corresponding to multiple consecutive past operating cycles into the target deep learning time series prediction model to obtain the remaining service life of the rail transit signal equipment output by the target deep learning time series prediction model. The target deep learning time series prediction model is obtained by training an initial deep learning time series prediction model based on a health index time series containing multiple historical operating cycles and a remaining lifespan label corresponding to the health index time series.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the data processing method based on rail transit signaling equipment as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the data processing method based on rail transit signaling equipment as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the data processing method based on rail transit signaling equipment as described in any one of claims 1 to 6.

Images